I would like to seek your kind advise and feedback on my regressions as I am not sure whether I am doing it correctly.
An introduction to my problem, I am looking to analyse the impact of COVID across different industries, for my data, I have 19 sub-sectors which encompasses 5 main sectors. I have generated a dummy year2020 to capture the COVID year.
I am using fe and have used:
Code:
xtset econsector
Code:
xtreg lnemp i.year i.quarter i.year2020##i.econsector, fe baselevels
note: 1.year2020 omitted because of collinearity
note: 2.econsector omitted because of collinearity
note: 3.econsector omitted because of collinearity
note: 4.econsector omitted because of collinearity
note: 5.econsector omitted because of collinearity
note: 6.econsector omitted because of collinearity
note: 7.econsector omitted because of collinearity
note: 8.econsector omitted because of collinearity
note: 9.econsector omitted because of collinearity
note: 10.econsector omitted because of collinearity
note: 11.econsector omitted because of collinearity
note: 12.econsector omitted because of collinearity
note: 13.econsector omitted because of collinearity
note: 14.econsector omitted because of collinearity
note: 15.econsector omitted because of collinearity
note: 16.econsector omitted because of collinearity
note: 17.econsector omitted because of collinearity
note: 18.econsector omitted because of collinearity
note: 19.econsector omitted because of collinearity
Fixed-effects (within) regression Number of obs = 456
Group variable: econsector Number of groups = 19
R-sq: Obs per group:
within = 0.4989 min = 24
between = 0.0978 avg = 24.0
overall = 0.0110 max = 24
F(26,411) = 15.74
corr(u_i, Xb) = 0.0763 Prob > F = 0.0000
-------------------------------------------------------------------------------------------------------------------------------------------
lnemp | Coef. Std. Err. t P>|t| [95% Conf. Interval]
--------------------------------------------------------------------------+----------------------------------------------------------------
year |
2015 | 0 (base)
2016 | .003923 .0058952 0.67 0.506 -.0076654 .0155115
2017 | .0270333 .0058952 4.59 0.000 .0154448 .0386217
2018 | .0488873 .0058952 8.29 0.000 .0372988 .0604757
2019 | .0708191 .0058952 12.01 0.000 .0592307 .0824076
2020 | .0437246 .0202506 2.16 0.031 .003917 .0835323
|
quarter |
1 | 0 (base)
2 | -.0064956 .0048134 -1.35 0.178 -.0159576 .0029663
3 | .000597 .0048134 0.12 0.901 -.0088649 .0100589
4 | .0109844 .0048134 2.28 0.023 .0015224 .0204463
|
year2020 |
0 | 0 (base)
1 | 0 (omitted)
|
econsector |
Agriculture | 0 (base)
Beverages and tobacco products | 0 (omitted)
Construction | 0 (omitted)
Electrical, electronic and optical products | 0 (omitted)
Finance and insurance | 0 (omitted)
Food & beverages and accommodation | 0 (omitted)
Information and communication | 0 (omitted)
Mining and quarrying | 0 (omitted)
Non-metallic mineral products, basic metal and fabricated metal products | 0 (omitted)
Other Services | 0 (omitted)
Petroleum, chemical, rubber and plastic products | 0 (omitted)
Real estate and business services | 0 (omitted)
Textiles, wearing apparel and leather products | 0 (omitted)
Transport equipment, other manufacturing and repair | 0 (omitted)
Transportation and storage | 0 (omitted)
Utilities | 0 (omitted)
Vegetable and animal oils & fats and food processing | 0 (omitted)
Wholesale and retail trade | 0 (omitted)
Wood products, furniture, paper products and printing | 0 (omitted)
|
year2020#econsector |
1#Beverages and tobacco products | -.0538645 .0281491 -1.91 0.056 -.1091986 .0014696
1#Construction | -.0813015 .0281491 -2.89 0.004 -.1366356 -.0259673
1#Electrical, electronic and optical products | .0318838 .0281491 1.13 0.258 -.0234504 .0872179
1#Finance and insurance | .0284702 .0281491 1.01 0.312 -.0268639 .0838044
1#Food & beverages and accommodation | .1063291 .0281491 3.78 0.000 .050995 .1616632
1#Information and communication | .1041067 .0281491 3.70 0.000 .0487726 .1594408
1#Mining and quarrying | -.0607854 .0281491 -2.16 0.031 -.1161196 -.0054513
1 #|
Non-metallic mineral products, basic metal and fabricated metal products | .0315135 .0281491 1.12 0.264 -.0238207 .0868476
1#Other Services | -.0031687 .0281491 -0.11 0.910 -.0585028 .0521655
1#Petroleum, chemical, rubber and plastic products | .038587 .0281491 1.37 0.171 -.0167472 .0939211
1#Real estate and business services | .0623288 .0281491 2.21 0.027 .0069947 .1176629
1#Textiles, wearing apparel and leather products | .0594772 .0281491 2.11 0.035 .0041431 .1148113
1#Transport equipment, other manufacturing and repair | -.042689 .0281491 -1.52 0.130 -.0980231 .0126451
1#Transportation and storage | .0281065 .0281491 1.00 0.319 -.0272276 .0834406
1#Utilities | .0254008 .0281491 0.90 0.367 -.0299334 .0807349
1#Vegetable and animal oils & fats and food processing | .062471 .0281491 2.22 0.027 .0071369 .1178051
1#Wholesale and retail trade | .0717998 .0281491 2.55 0.011 .0164657 .127134
1#Wood products, furniture, paper products and printing | -.0374862 .0281491 -1.33 0.184 -.0928203 .0178479
|
_cons | 6.002973 .0051054 1175.82 0.000 5.992937 6.013009
--------------------------------------------------------------------------+----------------------------------------------------------------
sigma_u | 1.2651704
sigma_e | .0363403
rho | .99917563 (fraction of variance due to u_i)
-------------------------------------------------------------------------------------------------------------------------------------------
F test that all u_i=0: F(18, 411) = 24241.02 Prob > F = 0.0000Code:
gen manuc = 0 replace manuc = 1 if econsector == 17 | econsector == 2 | econsector == 13 | econsector == 19 | econsector == 11 | econsector == 9 | econsector == 4 | econsector == 14 label var manuc "Manufacturing Sector"
Code:
xtreg lnemp i.year i.year2020##manuc, fe allbaselevels
note: 1.year2020 omitted because of collinearity
note: 1.manuc omitted because of collinearity
Fixed-effects (within) regression Number of obs = 456
Group variable: econsector Number of groups = 19
R-sq: Obs per group:
within = 0.3205 min = 24
between = 0.1706 avg = 24.0
overall = 0.0016 max = 24
F(6,431) = 33.88
corr(u_i, Xb) = 0.0176 Prob > F = 0.0000
--------------------------------------------------------------------------------
lnemp | Coef. Std. Err. t P>|t| [95% Conf. Interval]
---------------+----------------------------------------------------------------
year |
2015 | 0 (base)
2016 | .003923 .0067036 0.59 0.559 -.0092529 .0170989
2017 | .0270333 .0067036 4.03 0.000 .0138574 .0402092
2018 | .0488873 .0067036 7.29 0.000 .0357114 .0620631
2019 | .0708191 .0067036 10.56 0.000 .0576433 .083995
2020 | .0692961 .0080342 8.63 0.000 .053505 .0850872
|
year2020 |
0 | 0 (base)
1 | 0 (omitted)
|
manuc |
0 | 0 (base)
1 | 0 (omitted)
|
year2020#manuc |
0 0 | 0 (base)
0 1 | 0 (base)
1 0 | 0 (base)
1 1 | -.0143349 .0105172 -1.36 0.174 -.0350062 .0063364
|
_cons | 6.004244 .0047402 1266.67 0.000 5.994928 6.013561
---------------+----------------------------------------------------------------
sigma_u | 1.2674848
sigma_e | .041324
rho | .99893816 (fraction of variance due to u_i)
--------------------------------------------------------------------------------
F test that all u_i=0: F(18, 431) = 21938.97 Prob > F = 0.0000Do you think the second regression is feasible? or I should just use the first regression. Appreciate further comments on this.
Thank you!
0 Response to How to analyse dummy industry panel data
Post a Comment